Choose your query mode

IARA is a next-generation in silico methodology (NAM) that combines artificial intelligence with a Biomedical Knowledge Graph (BKG) to predict liver, cardiac, nervous system, and renal toxicity.

Select one of the two available query modes to perform the toxicity prediction.

Mode A

Compound identifier-based

Send up to 10 ChEMBL IDs for the compounds you want to query.

Accepted format: ChEMBL ID. One identifier per line.

Process compound identifier

Mode B

By protein target set

Send the protein targets of one compound. Add an optional confidence score from 0 to 10 on each line.

Format: UniProt ID, then an optional score.

Process protein set

How it works

01

Define your query

Enter the chemical identifier or the list of protein targets associated with the compound of interest.

02

IARA traverses the graph

The system navigates the curated human biomedical knowledge graph to identify relevant relationships.

03

Explainable result

Get organ-level toxicity predictions with a probability and an applicability-domain check.

First time with IARA?

Check how a query is built, which organs are evaluated, and how an out-of-domain result changes the reliability of a prediction.

See Documentation